Cultural entrainment of motor skill development: Learning to write hiragana in Japanese primary school

Abstract The aim of the present study was to examine how the social norms shared in a classroom environment influence the development of movement dynamics of handwriting of children who participate in the environment. To look into this issue, the following aspects of the entire period of classroom learning of hiragana letters in Japanese 1st graders who had just entered primary school were studied: First, the structure of classroom events and the specific types of interaction and learning within such environment were described. Second, in the experiment involving 6‐year‐old children who participated in the class, writing movements of children and their changes over the period of hiragana education were analyzed for each stroke composing letters. It was found that writing movement of children became differentiated in a manner specific to the different types of stroke endings, to which children were systematically encouraged to attend in the classroom. The results provide a detailed description of the process of how dynamics of fine motor movement of children is modulated by the social norms of a populated, classroom environment in a non‐Latin alphabet writing system.

Handwriting requires the control of continuous hand movements in such a way to leave a desired trajectory of the tip of the pen on a two-dimensional writing surface. The task, therefore, has traditionally been viewed as a task of controlling fine motor movements, different aspects of which have been investigated by a number of researchers (van Gemmert & Contreras-Vidal, 2015). For example, Greer and Lockman (1998) asked preschoolers to draw horizontal and vertical lines with pens at different locations on a page. These researchers found a reduction in variability both in grip pattern and pen-surface positioning between 3 and 5 years of age, and suggested that such developmental changes in the variability reflect the process of discovering forms of writing that maximize stability in orienting a tool to a surface such that certain spatial relations are maintained (Greer & Lockman, 1998). Newell and van Emmerik (1989) investigated the development of motor coordination in handwriting not in children but in adults, who practiced writing with their non-dominant hand their signature and cursive e's 1,000 and 10,000 times, respectively. The effect of practice turned out very small, implying that the acquisition of the skill may take much more practice than has been applied in experimental studies. In addition, it was found that in right-handers, the non-dominant limb links were more tightly frozen than was the case with the dominant limb, while this was not a case in left-handers, suggesting the influence of writing direction on the acquisition of coordination patterns (Newell & van Emmerik, 1989). More recently, Duval and co-workers (Duval, Rémi, Plamondon, Vaillant, & O'Reilly, 2015) studied the graphomotor skills of kindergarten children who produced pre-calligraphic trajectories, using different indexes of rapidity, fluidity, and regularity of their pen tip movements.
One of the curious facts about the previous studies on the development of handwriting is that the tasks performed by participants were often not the actual task of writing meaningful letters, but that of drawing or tracing of nonsense graphic forms such as straight lines and ellipses (Danna, Enderli, Athènes, & Zanone, 2012;Greer & Lockman, 1998), continuous lines of loop patterns (Bosga-Stork, Bosga, & Meulenbroek, 2011), or other abstract letter-like patterns (Duval et al., 2015;Jongbloed-Pereboom, Peeters, Overvelde, Nijhuis-van der Sanden, & Steenbergen, 2015). In these studies, the task of handwriting was viewed as the task of forming and producing a desired spatial trajectory of the tip of a tool held in hand. The development of the skill to plan and generate fine movements to precisely control the endpoint of a tool is, no doubt, an important topic of research. However, the research on the skills of tracing meaningless graphic forms inevitably leaves out an account of what it is that makes one's handwriting "meaningful," affording discrimination by others in a populated environment. And one is left to wonder whether it is possible to separate learning to control bodily movements involved in writing from learning to write in a socially meaningful manner.
In my view, the developmental research on handwriting needs to be complemented by something like the Vygotskian perspective (Vygotsky, 1978a) and the Gibsonian perspective (Gibson, E. J., 1971;Gibson J. J., 2015). Despite their different emphases, both Vygotsky and the Gibson's went beyond the physical/social dichotomy in important ways by placing primacy on the terrestrial, populated environment in which the development of human skills literally takes place. Vygotsky emphasized, among other things, the socially organized, future-oriented, yet nonlinear nature of child development (Vygotsky, 1978b). Human learning is prospective in a sense that it aims, in advance, for a new stage of the developmental process with the guidance of adults. Thereby, children grow into the life of those around them (Vygotsky, 1978b). Vygotsky, who viewed the development of writing as an interesting paradigm for the problem of the relation between learning and development, referred to the skill as "a complex cultural activity (Vygotsky, 1978c, p.118)" which develops along the path that it does because of cultural entrainment towards functioning in a populated environment. E. J. Gibson extensively studied how children learn to read (Gibson, 1965(Gibson, , 1970(Gibson, , 1971(Gibson, , 1975Gibson & Levin, 1975;Gibson, Gibson, Pick, & Osser, 1962). Through a series of experiments, Gibson showed that the ontogeny of reading in children involves progressive differentiation of the "distinctive features" of graphic information available in a given writing system (Gibson, 1971). Distinctive features are relational, not absolute like building blocks (Gibson, 1975). For example, 4-year olds may confuse Latin alphabet letters i, j, and t. Yet, adults may focus their attention to distinctive topological features such as how the two lines are laid out relative to one another, or curve-straight contrast at the end of a line to discriminate one letter from another, even if the letters are poorly handwritten and far from being similar or congruent to printed fonts. In other words, distinctive features are what remain invariant over a number of transformations that are irrelevant or noncritical for differentiating the symbols (Gibson, 1975). Likewise, in the act of writing, some relational features of traces left by the movement are expected to be critical and more tightly controlled than others, because these distinctive features must be made available in order for lines on paper to afford discrimination by other members of a language community. This view is in line with the argument presented by Bernstein from the perspective of motor control: "the human motor system cannot attain any high degree of metric proficiency, but it can be said that our motor system is very sensitive to topological distinctions of higher orders . . . . It is sufficient, for example, to draw attention to handwriting . . . that the letter A belongs to a single topological class of the first order no matter how or by whom it is written .. . . the analysis of which is not yet practicable for us because of our lack of acquaintance at present with whatever may constitute higher topological orders and what properties we must ascribe to them (Bernstein, 1984, pp.105-106)." If this line of reasoning is correct, then "what is learned" by children acquiring the skill of handwriting may not be so much the formation and reproduction of a graphic trajectory, but the ways to differentiate meaningful topological variations of lines by the movement of a writing implements on a writing surface. Arguably, addressing this point requires us to begin to take seriously the contexts of values in the populated environment to which the development of the skill progressively adapts (c.f., Bril, 1986;Gibson, 1950;Mauss, 1973;Tomasello, 1999;Vygotsky, 1978a).
What is important, but has not been studied, is how the populated environment entrains the perceiving and acting involved in handwriting skills of children. To explore this issue, the present study investigates the handwriting skill development of children in relation to the classroom environment in the beginning of handwriting education in Japanese primary school. By considering a syllabary that is unfamiliar to Latin alphabet communities, I aim to highlight and emphasize how the process of development of motor skills in handwriting is related to those dimensions that are critical in the populated environment where learning occurs.
The present study will concern a Japanese syllabary called hiragana. To add to the borrowed Chinese logograms, the Japanese developed two syllabaries: hiragana and katakana, each of which has 46 letters ( Figure 1a). It is possible to write Japanese using only either one of the syllabaries. But in practice, Japanese books are written with a combination of Chinese characters and the syllabaries, which makes them much easier to read than the syllabary alone because the Chinese characters serve to provide clues for dividing the text into word or meaning units (Sakamoto & Makita, 1973). Japanese prints texts from top-to-bottom from right-to-left, although both these vertical arrangements and the other horizontal, left-to-right printing are commonly used today. Each hiragana letter, as well as each Chinese character, is written according to a set sequence of strokes. To learn how to write a letter is to learn how to trace each of its elements in a precise order, which is itself determined by a certain number of rules. Among others, there are following three types of the ending of the stroke-how the element is closed-in hiragana ( Figure 1b): (1) "hooked" (brief stop followed by the abrupt reversal of pen tip movement at the end), (2) "sweep" (fast sweeping motion of the pen tip without change of direction), and (3) "stop" (slowing down to stop the pen tip to end the stroke). Every element in hiragana letters has either one of the above three types of ending, although there is variation in direction.
In Japan, prior to entering primary school, language education officially consists of verbal expression and listening (Ministry of Education, Culture, Sports, Science & Technology in Japan, 2011). It is not until the first months of the first grade when the official education of writing and reading begins. However, by the time the official education starts at the age of 6, most Japanese children have had some experience with writing implements and picture books, and have become able to write hiragana graphemes in their idiosyncratic manners more or less correctly in terms of their gross forms. The aim of the present study is to examine how such idiosyncratic patterns of movements involved in handwriting change when children are first exposed to a populated, classroom environment.
The main hypothesis of the present study was that the social norms shared in a classroom environment influence the development of movement dynamics of handwriting. Traditionally, cultural ecology of classroom learning and dynamics of handwriting movements have been studied and reported in different disciplines such as education and biomechanics (e.g., Graham et al., 2008;Shim et al., 2010).
Nevertheless, to address the issue of how the former entrains the latter, the analyses of both ecological context and individual development are inevitably required. The originality of the paper lies in the attempt at finding connection between these different aspects of the development of children. To test the hypothesis that the   year of primary school, approximately 100 class hours should be allocated to writing education (including the composition works), and that the 1st graders are expected to learn to be able "to read and write hiragana and katakana (Ministry of Education, Culture, Sports, Science & Technology in Japan, 2008)." All primary schools use one of the five official textbooks for Japanese language education. Although flexibility is given to schools and teachers to formulate curriculums, the variation in the design of classroom learning in Japanese primary schools is within a definite range reflecting the above requirements.

| Method
At Kobe University Elementary School where the present study was conducted, hiragana lessons lasted for 10 weeks, three to four class hours per week (Table 1)  During classroom practice, the aspects of the act of writing a letter frequently addressed by the teacher were classified into the following categories: (a) ending of the stroke (e.g., stop, sweep, or jump at the end); (b) curve/straightness of the stroke (e.g., make the stroke slightly curved); (c) turning of the stroke (e.g., make an acute angle); (d) intersection of the strokes (e.g., cross over the stroke); (e) proportions (e.g., make it shorter than the other element); (f) beginning (where to begin the stroke); (g) direction/slant of the stroke (the direction of the stroke or how it is to be slanted); (h) order of the strokes; and (i) rhythm (e.g., slowly here).
The author and two other coders (Japanese graduate students majoring psychology) coded all the comments and responses made by the teacher independently for each classroom activity. The intra-class correlation coefficients across the three raters were above .94 in all the segments of classroom practice coded, indicating high inter-rater reliability.
In what follows, I will report on the characteristics of classroom interaction observed in each of the above classroom activities.
1 Demonstration on the blackboard: Without exception, lessons started with a teacher's demonstration of writing the letter-to-be-learned on a square sheet attached on the blackboard. The sides of the square was about 30 cm long, and the square is ruled in such a way to divide the square into four small "chambers," so as to make it easy for children to know the relative position of each element composing a letter. Children were encouraged to raise their hands and trace the strokes rhythmically in the air with broad gestures of arm and hand, closely following the teacher's writing movement on the blackboard.
As the teacher writes on the blackboard, she would draw children's attention to the aspects of writing such as from which "chamber"   When a child was picked by the teacher and commented in front of the class, the teacher was never observed to judge whether the comment is correct or not. Instead, the teacher typically responded by exaggerating the feature commented on by each child in a manner contrary to that suggested by the child on the blackboard.
For example, when a child said "one should 'sweep' at the end of this stroke," the teacher would respond by writing the same stroke with a "hooked" ending on the blackboard exaggerating how improper the letter would look with a stroke ending in an incorrect manner, saying, "so you mean, you are not supposed to write like this." Such interaction was repeated with each children, and the "improper" variation of the letter in question accumulates on the blackboard, implying the variation of the letter that could not be tolerated  Teacher corrected children's activity in response to the class atmosphere (17%, often telling them to be quiet), their postures (12%), and the letters they had written (7%). This activity was usually the final part of the lesson, whose mean duration was 6 min 49 sthe longest among the five classroom activities.

| Timeline of instructional interventions
Despite the same sequence of classroom activities repeated for each letter, the instructional comments provided by the teacher exhibited both change and nonchange over the period (Figure 4). In the whole-

| STUDY 2: DEVELOPMENT OF HANDWRITING SKILLS
This part of the study aims to explore how the motor development of children using a stylus was modulated by the experience in this classroom environment. Since this is an exploratory study, I focused on general questions such as the following: Do the developmental

| Procedure
The trace-making movements of the stylus manipulated by each child when writing hiragana letters were recorded every other week. The first session was conducted in the 4th week of classroom hiragana learning, and the final session was done in the 12th week (i.e., recording sessions are done in 4th, 6th, 8th, 10th, 12th week-five sessions in total). Note that hiragana education lasted only for 10 weeks, and the 12th week was two weeks after the completion of hiragana letter learning classes (Table 1). Except one boy who missed the first recording session, all the participants completed all the five recording sessions. Each child was presented a sheet of ruled paper and was instructed to write each of two to three Japanese words for five times between two vertical lines whose inter-line distance was 1.4 cm. The model of the words to be written were printed on the righthand end of the same sheet of paper. The dimension of the model letter presented was approximately 0:8 Â 0:8 cm. The set of words to be written in each session was selected in such a way to always contain the same four hiragana letters which consisted of ten strokes in total.

| Data analysis
The horizontal and vertical position coordinates were low-pass filtered at 12 Hz using a sinusoidal transition band from 7 to 17 Hz using multiplication of the spectrum (Teulings & Maarse, 1984). The mutual information between a pair of variables X and Y was estimated using the histogram approach, in which the probability density function of each variable is approximated using a histogram (Fraser & Swinney, 1986). Then, the mutual information (MI) between a pair of variables X and Y can be calculated according to the following equation (Cover & Thomas, 1991): All measurements were calculated for each individual stroke for all participants (1347 strokes in total). A linear mixed-model analysis was used to test the effect of "week" and/or "type of stroke ending" on the means of each selected variable. To account for the correlation between strokes made by the same participant, a participant factor was included as an additional random-effect (Boyle & Willms, 2001).
Bonferroni-adjusted pairwise comparisons based on estimated marginal means were used for multiple comparisons between skill levels.
All statistical tests were made using SPSS 22.0. The alpha value for a significant effect was set at 0.05.

| Results and discussion
The modulation of handwriting movement during the period of official hiragana education were different across individuals, although there were also some common transformational invariants (c.f., Thelen et al., 1993). To capture this variability in developmental trajectory and to discover common strategies as well, the results section was divided into two sections: (1) collective results of all children and (2) examples of the developmental paths taken by individual children.

| Collective results
In the previous section, we have seen that children were systematically encouraged to attend to the movement with which one ends each stroke composing a letter. As mentioned previously, in contrast to "stop" ending, "hooked" and "sweep" endings involve fast sweeping motion of the pen tip at the very end of the stroke. at the end of the strokes with "hooked" and "sweep" endings were consistently greater compared to the strokes with "stop" ending (p < .05) (Figure 6a). Accompanying such divergence specific to different ending types, the overall increase of peak resultant speed over the course of two months was also observed ( Table 2).
While there was no consistent decrease of average duration it took for children to write one hiragana letter (Table 2)   The way in which children oriented a pen to a writing surface changed only a little. The children, who were all right handed, tilted the pen to their right-hand side from the beginning, the tendency of which seems to have gotten slightly more pronounced in the later sessions compared to the earlier sessions (Table 2). Apparently, there were no common, consistent changes across children in the size of the letters written, the total length of the pen tip trajectory of one stroke, the pressure at the pen tip, and the altitude of the pen in celestial coordinate system (Table 2).

| Examples of the paths of individual development
In what follows, detailed data on two girls who exhibited quite different developmental trajectories were selected and presented with the aim of highlighting the paths taken by different individuals underlying the collective results reported above. week. Similar tendency was also found in "sweep" ending (light-gray arrow in the bottom panel of Figure 8). The resultant speed profiles also changed. The second peak of the strokes with "hooked" ending became progressively shifted backwards in such a way to end at the peak of speed instead of its valley after the peak (Figure 7a). Overall, visual inspection of Figure 7 suggests that by the 12th week, the movement of the Yui had assumed consistent rhythmic patterns specific to the type of strokes. Reflecting the increased pen-tip speed at "hooked" ending, the written traces at the end of "hooked" strokes were thinner in the 12th week compared to the 4th week, and the forms of letters differed across these sessions in details of connections between strokes (Figure 9). The wobbliness of Yui's pen tip movement that were initially present gradually disappeared (Figure 7b,c), which was also shown by the decrease in the number of resultant speed minima in the later sessions compared to the earlier sessions ( Figure 10a).

The case of Ann
First recording The letters Ann wrote were very small in the beginning (Figure 13), her movements were relatively slow (Figure 11), and the pressure of the pen tip was also very low ( Figure 10b). In contrast to Yui whose pen tip movement always reached a halt at the end of the strokes in the earlier sessions, Ann's relatively small and slow movements already exhibited the rhythmic patterns specific to different types of strokes from the 4th week, such as a brief stop followed by the abrupt jump at the end in "hooked" endings ( Figure 11a). Her movements were a little wobbly when writing a winding stroke (the second row in Figure 11b,c), but overall, Ann exhibited mature-looking patterns of movements compared to the other children from the beginning.
Subsequent modulation Ann scaled up her pen tip velocity in the subsequent weeks ( Figure 11). Not only did Ann learn to produce higher pen tip velocity, Ann's movements were modulated in such a way to assume the pronounced specificity to the strokes with different types of ending ( Figure 12). Note the difference in the magnitude of velocity between "hooked" stroke (black arrows) and "stop" stroke (dark-gray arrows) in the upper panel in Figure 12. While Ann's pen tip movement reached an almost complete halt at the end of "stop" FIGURE 8 Progressive differentiation of the pen-tip movement of Yui between strokes with different endings composing hiragana letters. Black, dark-gray, and light-gray arrows on the right correspond to the ending velocities of pen-tip writing the elements with "hooked," "stop," and "sweep" endings of hiragana letter shown in black, dark-gray, and light-gray on the left, respectively. The size and direction of arrows show the magnitude and direction of average velocity within a 50 ms-window prior to lifting the pen. Overlaid arrows with the same color correspond to the multiple trials recorded at each week i) The patterns of movements in children writing hiragana letters became differentiated in a manner specific to the different types of stroke endings over the first months of primary school.
ii) Increased conformity of the temporal patterns of pen-tip movements over the period of classroom learning was observed within each individual, in such a way for a given letter to be written with a like temporal structure of movement.
iii) The developmental paths taken to differentiate movement patterns specific to the distinctive features of hiragana strokes were different across individual children.
iv) In the classroom, interactions between the teacher and children were often mediated by the ongoing "act" of writing, which involved shared rhythmic motor experience between them.
v) The lessons in the classroom were structured in such a way for children to see the interactive responses of the teacher to how they have acted, in which a consensus had grown as to which aspects of the act of writing matter.

| Transformations of pre-existing patterns of handwriting movements
One of the basic yet surprising outcomes of the present study was that Recently, Maldarelli and co-workers illustrated the development of visual attention of 5-year olds during letter copying task (Maldarelli, Kahrs, Hunt, & Lockman, 2015). The extent to which the visual attention develops in conjunction with education of attention in the classroom environment during such tasks may be an interesting topic of future research.
Another question of interest is how children come to be sensitive to such subtle distinctive features of elements composing hiragana. The classroom observation showed the characteristic manner in which the teacher directed children's attention to aspects of writing such as the order of the stroke and how the stroke should be closed. In the case of the skills such as reaching or pointing, how children move the hand through space to contact a target, and the consequences of their self-produced activity can be perceived with eyes, heads, mouths, and limbs through receptors in muscles, joints, skin, and vestibular system (Thelen et al., 1993).
Thereby, a movement repertoire of a child can be transformed into adaptive actions through the continual process of exploration of the perceptual consequences of self-generated movement (Sporns & Edelman, 1993 and 11 clearly illustrate that the rhythmic patterns of pen tip movements of children-how the pen tip lingers, leaps, and stops on a writing surface -have evolved in such a way to assume a consistent, stroke-specific temporal structure. In the collective data from all the six children, such specificity was exemplified by the increase of the mutual information between normalized resultant speeds between the same strokes within each child over multiple sessions (Figure 6c).
In the Japanese classroom where observations were made, learning of the letters were not cut off from their gestural foundations.
The form of each letter was regarded as a result of dynamic movement, such as how you end each stroke and its transitions to the following stroke. In some of the classroom activities, children first execute the gesture in the air with rhythmical movements which were coupled to that of the teacher writing on the blackboard, carefully making the endings. Each child was encouraged to attempt at moving one's own body in a particular temporal sequence, which may help the child resonate to the critical features that distinguish one letter from another. Such emphasis on gesture may be characteristic of societies under the influence of Chinese writing system (Billeter, 1990). Sasaki's (1987) series of experiments convincingly showed that when Japanese people cannot recall a Chinese character, they search with the hand until the gesture performs itself and restores the forgotten form, just as we would to recall a forgotten dance step. In fact, they are so accustomed to "mock writing" that when there is any uncertainty about a word when writing, they trace the character in the air or on their palm, and this significantly contributes to dispelling the uncertainty (Sasaki, 1987). Hiragana has its origin in Chinese character, and has characteristics in common in the set of strokes composing letters. It is not improbable that the process of learning letters that involves rhythmic motorial experience is related to this specific word recall strategies seen in Japanese people.
Traditionally, it has been assumed that underlying the skill of handwriting is the stored effector-independent visual representation of the strokes in handwriting (Wing, 2000), in which the two following tasks were said to exist: (1) the formation of the trajectory of the writing implement and (2) the production of the desired trajectory of the tip of the pen on a piece of paper (Latash, Danion, Scholz, Zatsiorsky, & Schöner, 2003;van Gemmert & Contreras-Vidal, 2015).
However, the results of the present study-that Japanese children acquire the specific temporal pattern of gesture corresponding to a stroke in the beginning of hiragana education-seem to suggest an alternative explanation. At least in this particular culture, what is learned may not be a picture of a letter plus the ability to trace that picture. Instead, it appears that learning of the temporal pattern of movement corresponding to a letter is primary, based on which the invariant features of a letter-the traces of the specific temporal pattern of movement-come to be discriminated as such. To ascertain meaningful invariants in the skill of handwriting across cultural variations, it would seem desirable that more studies on the development of handwriting in societies using non-Latin alphabet writing systems be conducted.
On the other hand, the increased specificity of temporal structure of the pen tip movement does not necessarily imply that what is acquired by the children were a set of muscle activation patterns. Just as the same melody could be sung or whistled, the same temporal structure of the movement on the pen tip on a horizontal piece of paper can emerge from a number of different coordination patterns among fingers that make contact with the pen (Latash et al., 2003;Shim et al., 2010). Nonaka (2013) reported on the motor coordination of a tetraplegic calligrapher who write with a writing implement gripped between his teeth. The study illustrated the movement variability of the head and that of the cervical spine were coupled in a compensatory manner to stabilize the upright posture of the brush relative to the contact with the paper, as the head flexed down and up, and rotated, to produce the same strokes in multiple trials (Nonaka, 2013). In the present study, although the data collected were restricted to the movement of the pen tip and the pen's orientation relative to a horizontal surface, the progressive changes in the movement of the pen tip found in children that accompanied only slight changes in pen orientation may be indicative of motor equivalent control of the pen tip. Underlying change in the strategies to control a redundant system to produce a specific pen tip movement may be a topic of future research.

| CONCLUSIONS
Traditionally, the development of handwriting has been studied as the